Questions: Phoneme Perception and Categorical Perception of Speech

Question 1

An English speaker is tested on discrimination of Hindi retroflex consonants that English does not distinguish phonemically. What would categorical perception theory predict?

• A. The English speaker will discriminate the Hindi contrast as well as a native Hindi speaker, because acoustic differences are universal
• B. The English speaker will show poor discrimination of the Hindi contrast because no phoneme boundary exists in English at that acoustic location
• C. The English speaker will learn to discriminate the contrast within minutes of exposure, demonstrating rapid plasticity
• D. The English speaker will hear the sounds as completely identical regardless of how large the acoustic difference is

Answer: (B)

Categorical perception means the auditory system is sharpened at native language phoneme boundaries and less sensitive elsewhere. Hindi draws a phoneme boundary between retroflex and dental consonants where English does not — English treats both as the same category. Without a boundary at that acoustic location, the English speaker cannot reliably discriminate between sounds that differ across that dimension. Option A incorrectly assumes acoustic sensitivity is universal and unaffected by language experience — this is exactly the misconception categorical perception theory refutes.

Question 2

A speech synthesizer creates a 10-step VOT continuum from -20ms to +80ms. An English listener identifies steps 1-5 as /b/ and steps 6-10 as /p/. Compared to discrimination of steps 3 vs. 4, discrimination of steps 5 vs. 6 will be:

• A. Worse, because the sounds near the boundary are the most acoustically ambiguous
• B. The same, since the acoustic distance (VOT change) is identical in both pairs
• C. Better, because steps 5 and 6 straddle the phoneme category boundary
• D. Impossible to predict without knowing the listener's language history

Answer: (C)

The defining characteristic of categorical perception is asymmetric discrimination: pairs that differ by the same acoustic distance are discriminated much better when they cross a phoneme boundary than when they fall within the same category. Steps 5 and 6 straddle the /b/-/p/ boundary, so even though the physical VOT difference is the same as between steps 3 and 4, discrimination is dramatically better at the boundary. Option B represents the naive prediction if perception were continuously proportional to acoustic difference — categorical perception violates this prediction, which is what makes it theoretically important.

Question 3

Infants under 6 months of age can discriminate phoneme contrasts from languages they have never been exposed to.

• T. True
• F. False

Answer: True

This foundational finding demonstrates that categorical perception is largely learned, not innate. Young infants possess universal phonetic sensitivity — they discriminate virtually any phoneme contrast in any human language, including click consonants and retroflex distinctions not present in their environment. Between 6-12 months, native language experience sculpts this sensitivity: contrasts used in the ambient language are sharpened while contrasts not used fade. By 12 months, infants show the adult pattern of categorical perception tuned to their native language — demonstrating that boundaries are acquired through statistical exposure, not predetermined by biology.

Question 4

Categorical perception of speech sounds reflects innate, language-universal acoustic sensitivity built into the human auditory system.

• T. True
• F. False

Answer: False

While there may be some initial biases, categorical perception is primarily learned through language experience. Phoneme boundaries occur at different acoustic values across languages — the English /b/-/p/ boundary falls at ~25ms VOT while the Spanish boundary falls at ~0ms — which cannot be explained by innate universal processing. Cross-linguistic evidence and infant developmental data both demonstrate that boundaries are sculpted by statistical exposure to the native language during a critical period. If categorical perception were purely innate, all humans would share identical phoneme boundaries regardless of which language they heard.

Question 5

Why do Japanese speakers have persistent difficulty distinguishing English /r/ from /l/, even after years of English exposure?

Model answer: Japanese does not draw a phoneme boundary at the acoustic location that separates English /r/ from /l/. During language acquisition, the Japanese speaker's perceptual system never established a category boundary at that point in acoustic space — both sounds fall within a single Japanese phonemic category. Without a boundary there, both sounds are perceived as instances of the same phoneme. The difficulty persists because adult phoneme categories are deeply entrenched; the critical period during which boundaries are most easily established has passed, and restructuring requires significant effortful training.

Categorical perception theory predicts that discrimination is sharpest at phoneme boundaries and poor within categories. When a second language draws boundaries at locations unmarked by the native language, adult learners must try to establish new categories against strongly established existing ones — an especially difficult task for fine acoustic distinctions that fall within a single native language category.